Reply To: cantilever beam problem

[wrbulat]

Adding to what Dave said, the magnetostructural coupled options of SOLID226/227 are not natively exposed in Mechanical, so you would have to use command objects to at least (a) call the option into use (set KEYOPTION 1 to 10001 or 10101), and (b) define magnetic material properties (relative permeability to region occupied by iron particles) and possibly electric conductivity to the aluminum (if motion induced eddy currents and corresponding Lorentz forces are significant). 

The presence of nonlinear (displacement dependent) forces is almost sure to force you to perform a transient analysis, with the beam initially deflected, then released. If the arrangement is such that the magnet(s) and the distribution of the forces they exert are relatively localized, you might be able to devise a lumped approximation of the magnetic force with spring elements (a less computationally expensive option). Springs can be defined in such a way as to produce nonlinear negative stiffness (with attractive forces increasing as nodes get closer) - please look at Negative stiffness in x-position between PM and steel stator (ansys.com) if interested.

I've never tried this myself, but you might be able to use the harmonic balance method in a harmonic response analysis to solve this problem (if the magnetic forces are simplified with nonlinear springs):

HBM seems a pretty steep learning curve to me - it might be simpler to start with a nonlinear transient analysis.

--Bill